GEOLOGY | Volume 44 | Number 6 | www.gsapubs.org 419

Insights into surface runoff on early Mars from paleolake basin morphology and stratigraphyTimothy A. Goudge1*, Caleb I. Fassett2, James W. Head1, John F. Mustard1, and Kelsey L. Aureli11Department of Earth, Environmental and Planetary Sciences, Brown University, Providence, Rhode Island 02912, USA2Department of Astronomy, Mount Holyoke College, South Hadley, Massachusetts 01075, USA

ABSTRACTWe present observations on the morphology and stratigraphy of more than 400 paleolake

basins on Mars. We show that there are two distinct classes of Martian paleolake basins: (1) paleolakes fed by regionally integrated valley networks (N = 251), and (2) paleolakes fed by isolated inlet valleys not integrated into broader regional drainage systems (N = 174). We conclude that valley network–fed paleolakes primarily formed prior to approximately the Noachian-Hesperian boundary, ca. 3.7 Ga, while isolated inlet valley paleolakes primarily formed later in Martian history. All 174 isolated inlet valley paleolakes are closed-basin lakes; however, there are surprisingly few (31) valley network–fed closed-basin lakes compared to a large number (220) of valley network–fed open-basin lakes. This observation is consistent with declining levels of fluvial activity over time on the Martian surface. Our results imply that during the era of valley network formation, ~90% of topographic basins breached by an inlet valley had sufficiently high ratios of water influx to losses to fill, overtop, and form an outlet valley. This conclusion provides an important constraint on the balance between surface runoff production and water losses on early Mars that must be satisfied by any model of the early Martian climate and hydrologic cycle.

INTRODUCTIONThe southern highlands of Mars contain abun-

dant geomorphic evidence for fluvial activity, largely associated with valley network forma-tion that primarily ceased around the Noachian-Hesperian boundary, ca. 3.7 Ga (e.g., Carr, 1996; Howard et al., 2005; Irwin et al., 2005; Fassett and Head, 2008a; Hynek et al., 2010; Mangold et al., 2012). Paleolake basins are key features that indicate past fluvial activity (e.g., Forsythe and Blackwelder, 1998; Cabrol and Grin, 1999, 2010; Irwin et al., 2005; Fassett and Head, 2008b; Goudge et al., 2012, 2015). Catalogs of Martian paleolakes have been made for both hydrologically closed and hydrologically open basins (Fassett and Head, 2008b; Goudge et al., 2012, 2015).

Closed-basin lakes in the Goudge et al. (2015) catalog have an inlet valley breaching the basin-confining topography, but no outlet valley (Figs. 1A and 1B). These features are best referred to as candidate paleolakes, as the presence of a valley flowing into a basin does not require the formation of a lake; however, the for-mation of an inlet valley requires fluvial incision and transport of water into the basin, so for sim-plicity we refer to these basins as closed-basin lakes. Open-basin lakes are typically fed by inlet valleys and have an outlet valley that breaches the basin-confining topography (Fig. 1C), which requires that the basin was once filled with water

prior to incising an outlet breach (Fassett and Head, 2008b).

While both closed- and open-basin lakes have been widely documented on Mars (Cabrol and Grin, 1999, 2010; Fassett and Head, 2008b; Goudge et al., 2012, 2015), the relationship between these two sets of features is unclear. Did closed- and open-basin lakes form during the same era of Martian history? Did they involve similar levels of fluvial activity? What can these basins tell us about the surface environment of early Mars? Here we analyze the morphology and stratigraphy of a catalog of more than 400 paleo lakes to provide insight into these questions.

DATA SETS AND METHODSWe studied the morphology, degradation

state, and stratigraphy of 205 closed-basin lakes (e.g., Figs. 1A and 1B; Goudge et al., 2015) and 220 open-basin lakes (e.g., Fig. 1C; Fassett and Head, 2008b; Goudge et al., 2012), along with their associated fluvial valleys, using image and topographic data. The primary image data sets include ~6 m/pixel images from the Context Camera (CTX; Malin et al., 2007), <50 m/pixel images from the High Resolution Stereo Cam-era (HRSC; Neukum et al., 2004), and the ~100 m/pixel global daytime infrared mosaic (Edwards et al., 2011) from the Thermal Emis-sion Imaging System (THEMIS; Christensen et al., 2004). For topographic analyses, we used

~463 m/pixel global gridded topography from the Mars Orbiter Laser Altimeter (MOLA; Smith et al., 2001).

The morphology of inlet and outlet valleys associated with the catalog of paleolakes was assessed based on the level of regional fluvial integration and incision. Particular attention was paid to comparing the morphology of these valleys with the morphology of ancient valley networks, which have multiple branching tribu-taries and are integrated with the surrounding landscape (e.g., Howard et al., 2005; Irwin et al., 2005; Hynek et al., 2010; see the GSA Data Repository1 for additional details).

We analyzed the stratigraphy and degradation state of the basins that host the >400 paleolakes, which are largely defined by impact craters (Fassett and Head, 2008b; Goudge et al., 2012, 2015), to provide an estimate of the timing of paleolake activity (for additional details, see the Data Repository). Crater degradation is a well-studied process on Mars, and older craters tend to be more degraded than younger craters of similar diameters (e.g., Craddock et al., 1997; Craddock and Howard, 2002; Mangold et al., 2012; Rob-bins and Hynek, 2012). Primary effects of crater degradation include basin infilling, removal of the central peak or pit, erosion and modification of ejecta deposits, and backwasting and erosion of crater rims and walls (Craddock et al., 1997; Mangold et al., 2012).

RESULTS

Paleolake Valley MorphologyWe identify 31 closed-basin lakes fed by

inlet valleys with a morphology consistent with typical Martian valley networks (Howard et al., 2005; Irwin et al., 2005; Hynek et al., 2010), i.e., regionally integrated valleys with mul-tiple branching tributaries and Strahler orders commonly ≥2 (e.g., Fig. 1B; Figs. DR1 and

1 GSA Data Repository item 2016136, additional detail on methods and results, Figures DR1–DR5, and Tables DR1 and DR2, is available online at www .geosociety .org /pubs /ft2016.htm, or on request from editing @geosociety.org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.

*Current address: Jackson School of Geosciences, University of Texas at Austin, Austin, Texas 78712-1722, USA; E-mail: tgoudge@jsg.utexas.edu.

GEOLOGY, June 2016; v. 44; no. 6; p. 419–422 | Data Repository item 2016136 | doi:10.1130/G37734.1 | Published online 19 April 2016

© 2016 Geological Society of America. For permission to copy, contact editing@geosociety.org.

420 www.gsapubs.org | Volume 44 | Number 6 | GEOLOGY

DR5A; Table DR1; see the Data Repository). The remaining 174 closed-basin lakes are fed by isolated inlet valleys, which have few or no tributaries (i.e., typical Strahler orders of 1) and are poorly integrated with the surrounding landscape (e.g., Fig. 1A; Figs. DR2 and DR5A; Table DR1). All 220 open-basin lakes are asso-ciated with inlet and outlet valleys that have morphologies consistent with typical Martian valley networks (e.g., Fig. 1C; Figs. DR3 and DR5A; Table DR1; Howard et al., 2005; Irwin et al., 2005; Fassett and Head, 2008b; Hynek et al., 2010). We collectively refer to the group of 31 closed-basin lakes and 220 open-basin lakes associated with valley networks as valley net-work–fed paleolakes, while the remaining 174 closed-basin lakes are referred to as isolated inlet valley paleolakes.

Paleolake Degradation State and Stratigraphy

Valley network–fed paleolakes are typically contained within heavily degraded impact cra-ters with eroded rims and walls and infilled cra-ter floors (e.g., Figs. 1B and 1C; Figs. DR1, DR3, and DR5B). Many isolated inlet valley paleolakes, however, are contained within less degraded impact craters that often show ter-raced walls and central uplift structures (e.g., Fig. 1A; Figs. DR2 and DR5B). We also iden-tify 44 paleo lakes hosted by craters with pre-served continuous ejecta deposits or with inlet valleys that incise continuous ejecta deposits from nearby craters (Figs. 2 and 3; Fig. DR4; Table DR2). Of these, 41 basins host isolated

inlet valley paleolakes, and three host valley net-work–fed paleolakes, all of which are closed-basin lakes (Fig. 3; Table DR2). We also find two examples of isolated inlet valley closed-basin lakes hosted by impact craters with preserved continuous ejecta deposits that clearly superpose nearby valley networks (e.g., Fig. 2C).

IMPLICATIONS FOR PALEOLAKE ACTIVITY

Timing of Paleolake ActivityBased on the degradation state of the host

impact craters and the morphology of the associ-ated valleys, we conclude that the formation of valley network–fed paleolakes occurred during the era of major valley network activity, which primarily ceased ca. 3.7 Ga, around the time of the Noachian-Hesperian boundary (Howard et al., 2005; Irwin et al., 2005; Fassett and Head, 2008a, 2008b; Mangold et al., 2012). Therefore, valley network–fed paleolake basins represent the paleolake record from early Mars.

In contrast, we conclude that isolated inlet valley paleolakes primarily formed subsequent to the era of major valley network formation, based on three main observations. First, the host craters for isolated inlet valley paleolakes are morphologically fresher (e.g., Fig. 1A; Figs. DR2 and DR5B), suggesting a younger age of formation compared to the more heav-ily degraded craters hosting valley network–fed paleolakes (e.g., Figs. 1B and 1C; Figs. DR1, DR3, and DR5B). Second, 41 isolated inlet val-ley paleolakes are hosted by craters with intact

continuous ejecta deposits or have inlet valleys crosscutting the continuous ejecta deposits of nearby craters (Figs. 2 and 3; Fig. DR4; Table DR2), again suggesting a geologically recent age of formation (e.g., Mangold et al., 2012). Finally, in two locations, stratigraphic relation-ships show that the host craters with isolated inlet valleys are superposed upon ancient valley network systems (e.g., Fig. 2C), requiring that the paleolake formed after the preexisting valley network ceased incision.

Average Rates of Surface Runoff Production versus Water Losses

Our results also suggest differences in the average rates of surface runoff production ver-sus water losses (e.g., via infiltration or evapo-ration) associated with the formation of valley network–fed paleolakes and isolated inlet val-ley paleolakes. We note that the small number of valley network–fed closed-basin lakes (31) is in stark contrast to the large number of val-ley network–fed open-basin lakes (220). This requires that during the era of valley network formation, global average rates of surface run-off production versus water losses must have been sufficient for ~90% of craters breached by an inlet valley to fill with standing water, overtop, and incise an outlet valley to form a stable open-basin lake, as opposed to forming a hydrologically stable closed-basin lake. Isolated inlet valley paleolakes, however, are all closed-basin lakes, consistent with our conclusion that these basins formed later in Martian history, sub-sequent to the major period of valley network

CBA

Figure 1. Examples of paleolakes on Mars. White arrows indicate inlet valleys and red arrows indicate outlet valleys. A: Isolated inlet valley closed-basin lake at -9.6°N, 144.1°E. Image shows Mars Orbiter Laser Altimeter (MOLA) gridded topography overlaid on a mosaic of Context Camera (CTX) images P16_007158_1703 and B02_010230_1715. North is to the right. B: Valley network–fed closed-basin lake at -18.8°N, 59.2°E. Image shows MOLA gridded topography overlaid on a mosaic of CTX images B21_017868_1583 and P16_007148_1628, and High Resolution Stereo Camera (HRSC) nadir images h8440_0000 and h0532_0000. North is to the right. C: Valley network–fed open-basin lake at -20.5°N, 87.0°E. Image shows MOLA gridded topography overlaid on a mosaic of CTX images B05_011498_1584, B10_013621_1594, G08_021480_1594, and G14_023748_1601 and HRSC nadir image h6523_0000. North is to the right.

GEOLOGY | Volume 44 | Number 6 | www.gsapubs.org 421

activity. We hypothesize that these younger basins were precluded from filling and over-topping due to (1) less intense surface runoff production, (2) faster rates of water loss, and/or (3) a shorter overall time period of associated fluvial activity.

Eras of Martian Paleolake ActivityOur conclusions indicate two distinct eras

of Martian paleolake formation and associated climate conditions. The first era of paleolake activity was associated with the formation of valley network–fed, primarily open-basin lakes driven by a climate with higher surface runoff production compared with water losses, and occurred during the period of major valley net-work formation. The second era, which occurred later in Martian history, was associated with the formation of closed-basin lakes fed by isolated inlet valleys, with a climate characterized by a lower ratio of surface runoff production to water losses and/or shorter overall periods of fluvial activity. This conclusion is consistent with a hydrologic evolution of Mars with relatively intense fluvial activity in a period that ceased at approximately the Noachian-Hesperian bound-ary, and only limited fluvial activity during the Hesperian and Amazonian (e.g., Howard et al., 2005; Irwin et al., 2005; Fassett and Head, 2011; Mangold et al., 2012).

SURFACE RUNOFF ON EARLY MARSOur analysis provides an important broad

constraint on Martian hydrology and climate during the period of valley network formation: time-averaged surface runoff production com-pared with water losses during this era must have been sufficiently large to fill and overtop ~90% of paleolake basins breached by an incised inlet valley. We note that our analyses do not con-sider potential paleolakes not fed by an inlet

valley, such as enclosed basins with shallow lakes in their interiors fed by internal drain-age, e.g., Columbus crater (Wray et al., 2011). Such paleolakes may have been prominent on the early Martian surface; however, geomorphic evidence for standing water in closed basins without external tributaries is typically lack-ing, which makes identification of any possible paleolakes in this class difficult.

For paleolake basins on early Mars that were breached by an inlet valley, a strong control on whether the basin filled and breached to form an outlet valley would have been the ratio of watershed area to lake area. This ratio controls the relative amount of surface runoff-derived input a basin receives compared to the amount of water the basin loses through evaporation off the lake surface. All else being equal, basins on early Mars with larger watershed area to lake area ratios would have been more likely to fill and breach an outlet valley (e.g., Fassett and Head, 2008b).

An additional controlling factor on the for-mation of open-basin versus closed-basin lakes on early Mars would have been basin geogra-phy, as the early Martian climate is likely to have had significant spatial variation (e.g., Urata and Toon, 2013). Certain regions of Mars are likely to have had higher (or lower) than aver-age surface runoff production compared with water losses. We note a marked paucity of valley network–fed closed-basin lakes compared with valley network–fed open-basin lakes in the Mar-garitifer Sinus region, from ~-60° to 0°N and ~-30° to 30°E (Fig. 3). This may suggest that Margaritifer Sinus had a notably high ratio of surface runoff production to water losses, con-sistent with the observation that this region is one of the most densely dissected areas of Mars (e.g., Hynek and Phillips, 2001; Craddock and Howard, 2002; Grant and Parker, 2002; Howard

C

B

A

Figure 2. Examples of closed-basin lakes with isolated inlet valleys (white arrows) hosted by impact craters with continuous ejecta depos-its (orange arrows). A: Closed-basin lake at 27.7°N, -124.5°E. Image shows Mars Orbiter Laser Altimeter (MOLA) gridded topography overlaid on a mosaic of Context Camera (CTX) images P14_006693_2101, B08_012600_2072, and G19_025628_2068 and the Thermal Emis-sion Imaging System (THEMIS) ~100 m/pixel global daytime infrared mosaic. North is down. B: Closed-basin lake at 27.1°N, 21.4°E. Image shows MOLA gridded topography overlaid on a mosaic of High Resolution Stereo Camera (HRSC) nadir images h7286_0000, h7311_0000, and h9579_0001 and the THEMIS ~100 m/pixel global daytime infrared mosaic. North is up. C: Closed-basin lake at 31.4°N, -12.9°E. Note that the continuous ejecta deposit of the host crater (black outline) superposes a nearby valley network (blue arrows). Image shows MOLA gridded topography overlaid on the THEMIS ~100 m/pixel global daytime infrared mosaic. North is up.

Figure 3. Distribution of paleolakes on Mars. Valley network–fed open-basin lakes (white) and closed-basin lakes (black), and isolated inlet valley closed-basin lakes (red). Closed-basin lakes indicated by stars are hosted by craters with (w/) continuous ejecta deposits or have inlet valleys that incise continuous ejecta deposits of nearby craters. Background is Mars Orbiter Laser Altimeter (MOLA) gridded topography overlain on MOLA-derived hillshade map.

422 www.gsapubs.org | Volume 44 | Number 6 | GEOLOGY

et al., 2005; Barnhart et al., 2009; Hynek et al., 2010) and has a high concentration of large, inte-grated lake-chain systems (Fassett and Head, 2008b).

Our results indicate that the ratio of surface runoff production to water losses on the early Martian surface was both high and spatially vari-able; however, they do not constrain the total length of time over which valley networks were active. Both a short-lived period of valley net-work incision with large surface runoff fluxes and a longer lived period with lower surface run-off fluxes are consistent with our observations.

In a scenario with short-lived, perhaps cata-strophic, valley network activity where surface runoff production greatly exceeds water losses, the pre-breach volumes of valley network–fed open-basin lakes provide a minimum constraint on the amount of surface runoff that must be generated from the watersheds of those basins. For open-basin lakes hypothesized in Fassett and Head (2008b) to be mostly fed by surface runoff derived from their catchment, this typi-cally corresponds to meters to tens of meters of water spread across the entire watershed.

Alternatively, in a scenario with longer lived valley network activity and water loss rates that are comparable to surface runoff fluxes, our results require that for a typical paleolake that formed in this era, time-averaged water influx was greater than time-averaged water losses. A valley network–forming era characterized by multiple short episodes of fluvial activity insuf-ficient to cause widespread outlet breaching and separated by long periods of drought and evapo-ration is inconsistent with our observations: such a scenario would have resulted in far more valley network–fed closed-basin lakes than observed.

CONCLUSIONSAnalysis of the morphology, degradation

state, and stratigraphy of more than 400 paleo-lakes on Mars has revealed two distinct classes of basins: (1) valley network–fed paleolakes (e.g., Figs. 1B and 1C) that formed prior to ca. 3.7 Ga, and (2) isolated inlet valley paleolakes (e.g., Fig. 1A) that formed subsequently. Our observations are consistent with secular changes in the Mars environment and declining levels of fluvial activity throughout Martian history. Our analysis of the Martian paleolake record also reveals a fundamental characteristic of the hydrology of early Mars: ~90% of paleolake basins breached by well-incised inlet valleys were hydrologically open, requiring significant amounts of water influx to completely fill and overtop the vast majority of these basins. Future climate models must be able to sufficiently explain surface runoff production on early Mars that is consistent with this constraint.

ACKNOWLEDGMENTSWe thank J.L. Dickson for help with image and topo-graphic data processing, and K.E. Scanlon for helpful discussions. We also thank R.P. Irwin, R.A. Craddock, B.M. Hynek, and an anonymous reviewer for careful reviews that greatly improved the quality of this manu-script, and J.B. Murphy for editorial handling. Goudge gratefully acknowledges support for this work from the Natural Sciences and Engineering Research Council of Canada (NSERC) Postgraduate Scholarships Program (PGSD3–421594–2012). Head acknowledges support from membership on the ESA Mars Express HRSC Team through grant JPL-1488322, and the Mars Data Analysis Program through grant NNX11AI81G.

REFERENCES CITEDBarnhart, C.J., Howard, A.D., and Moore, J.M., 2009,

Long-term precipitation and late-stage valley net-work formation: Landform simulations of Parana Basin, Mars: Journal of Geophysical Research, v. 114, E01003, doi: 10 .1029 /2008JE003122.

Cabrol, N.A., and Grin, E.A., 1999, Distribution, clas-sification, and ages of Martian impact crater lakes: Icarus, v. 142, p. 160–172, doi: 10 .1006 /icar .1999 .6191.

Cabrol, N.A., and Grin, E.A., 2010, Searching for lakes on Mars: Four decades of exploration, in Cabrol, N.A., and Grin, E.A., eds., Lakes on Mars: Ox-ford, UK, Elsevier, p. 1–29, doi: 10 .1016 /B978 -0 -444 -52854 -4 .00001 -5.

Carr, M.H., 1996, Water on Mars: New York, Oxford University Press, 229 p.

Christensen, P.R., et al., 2004, The Thermal Emission Imaging System (THEMIS) for the Mars 2001 Odyssey mission: Space Science Reviews, v. 110, p. 85–130, doi: 10 .1023 /B: SPAC .0000021008 .16305 .94.

Craddock, R.A., and Howard, A.D., 2002, The case for rainfall on a warm, wet early Mars: Journal of Geophysical Research, v. 107, p. 21-1–21-36, doi: 10 .1029 /2001JE001505.

Craddock, R.A., Maxwell, T.A., and Howard, A.D., 1997, Crater morphometry and modification in the Sinus Sabaeus and Margaritifer Sinus regions of Mars: Journal of Geophysical Research, v. 102, p. 13,321–13,340, doi: 10 .1029 /97JE01084.

Edwards, C.S., Nowicki, K.J., Christensen, P.R., Hill, J., Gorelick, N., and Murray, K., 2011, Mo saick ing of global planetary image datasets: 1. Techniques and data processing for Thermal Emission Imaging System (THEMIS) multi-spectral data: Journal of Geophysical Research, v. 116, E10008, doi: 10 .1029 /2010JE003755.

Fassett, C.I., and Head, J.W., 2008a, The timing of martian valley network activity: Constraints from buffered crater counting: Icarus, v. 195, p. 61–89, doi: 10 .1016 /j .icarus .2007 .12 .009.

Fassett, C.I., and Head, J.W., 2008b, Valley network-fed, open-basin lakes on Mars: Distribution and implications for Noachian surface and subsurface hydrology: Icarus, v. 198, p. 37–56, doi: 10 .1016 /j .icarus .2008 .06 .016.

Fassett, C.I., and Head, J.W., 2011, Sequence and tim-ing of conditions on early Mars: Icarus, v. 211, p. 1204–1214, doi: 10 .1016 /j .icarus .2010 .11 .014.

Forsythe, R.D., and Blackwelder, C.B., 1998, Closed drainage crater basins of the Martian highlands: Constraints on the early Martian hydrologic cycle: Journal of Geophysical Research, v. 103, p. 31,421–31,431, doi: 10 .1029 /98JE01966.

Goudge, T.A., Head, J.W., Mustard, J.F., and Fassett, C.I., 2012, An analysis of open-basin lake depos-its on Mars: Evidence for the nature of associated

lacustrine deposits and post-lacustrine modifica-tion processes: Icarus, v. 219, p. 211–229, doi: 10 .1016 /j .icarus .2012 .02 .027.

Goudge, T.A., Aureli, K.L., Head, J.W., Fassett, C.I., and Mustard, J.F., 2015, Classification and analy-sis of candidate impact crater-hosted closed-basin lakes on Mars: Icarus, v. 260, p. 346–367, doi: 10 .1016 /j .icarus .2015 .07 .026.

Grant, J.A., and Parker, T.J., 2002, Drainage evolution in the Margaritifer Sinus region, Mars: Journal of Geophysical Research, v. 107, 5066, doi: 10 .1029 /2001JE001678.

Howard, A.D., Moore, J.M., and Irwin, R.P., 2005, An intense terminal epoch of widespread flu-vial activity on early Mars: 1. Valley network incision and associated deposits: Journal of Geo-physical Research, v. 110, E12S14, doi: 10 .1029 /2005JE002459.

Hynek, B.M., and Phillips, R.J., 2001, Evidence for extensive denudation of the Martian highlands: Geology, v. 29, p. 407–410, doi: 10 .1130 /0091 -7613 (2001)029 <0407: EFEDOT>2 .0 .CO;2.

Hynek, B.M., Beach, M., and Hoke, M.R.T., 2010, Up-dated global map of Martian valley networks and implications for climate and hydrologic processes: Journal of Geophysical Research, v. 115, E09008, doi: 10 .1029 /2009JE003548.

Irwin, R.P., Howard, A.D., Craddock, R.A., and Moore, J.M., 2005, An intense terminal epoch of wide-spread fluvial activity on early Mars: 2. Increased runoff and paleolake development: Journal of Geo physical Research, v. 110, E12S15, doi: 10 .1029 /2005JE002460.

Malin, M.C., et al., 2007, Context Camera investiga-tion on board the Mars Reconnaissance Orbiter: Journal of Geophysical Research, v. 112, E05S04, doi: 10 .1029 /2006JE002808.

Mangold, N., Adeli, S., Conway, S., Ansan, V., and Langlais, B., 2012, A chronology of early Mars climatic evolution from impact crater degrada-tion: Journal of Geophysical Research, v. 117, E04003, doi: 10 .1029 /2011JE004005.

Neukum, G., Jaumann, R., and the HRSC Co-Inves-tigator and Experiment Team, 2004, HRSC: The High Resolution Stereo Camera of Mars Express, in Wilson, A., ed., Mars Express: The scientific payload: European Space Agency Special Pub-lication SP-1240, p. 17–35.

Robbins, S.J., and Hynek, B.M., 2012, A new global database of Mars impact craters ≥1 km: 1. Data-base creation, properties, and parameters: Journal of Geophysical Research, v. 117, E05004, doi: 10 .1029 /2011JE003966.

Smith, D.E., et al., 2001, Mars Orbiter Laser Altim-eter: Experiment summary after the first year of global mapping of Mars: Journal of Geophysical Research, v. 106, p. 23,689–23,722, doi: 10 .1029 /2000JE001364.

Urata, R.A., and Toon, O.B., 2013, Simulations of the martian hydrologic cycle with a general circula-tion model: Implications for the ancient martian climate: Icarus, v. 226, p. 229–250, doi: 10 .1016 /j .icarus .2013 .05 .014.

Wray, J.J., et al., 2011, Columbus crater and other possible groundwater-fed paleolakes of Terra Si-renum, Mars: Journal of Geophysical Research, v. 116, E01001, doi: 10 .1029 /2010JE003694.

Manuscript received 27 January 2016 Revised manuscript received 5 April 2016 Manuscript accepted 6 April 2016

Printed in USA